Corruption : A Policy Perspective By Pranab Bardhan In many poor countries corruption is pervasive and widely presumed to be very hard to crack. In the face of this the common approach to the problem has often been either moralist ( only a fundamental change in moral values and standards in the society can resolve it ) or fatalist ( it s so bad that nothing can now be done ). The economists usually take an intermediate approach, indicating the need for a change in the incentive structure in such a way that the old adage honesty is the best policy becomes meaningful as a strategic goal, although some economists are aware that monetary incentives themselves can sometimes crowd out intrinsic motivations for moral behavior. There is a substantial theoretical literature now on the factors determining the incidence of corruption, suggesting some policies to fight it, but we do not have a great deal of empirical evidence on the effectiveness of these policies. In view of the inherent difficulty of carrying out such empirical studies on the impact of policies on corruption in the real world, there is now a growing experimental literature using simulation of corrupt transactions in controlled settings -- for a brief survey of this literature, see Abbink and Serra
(2012)--which is suggestive, though limited in its own way. This paper is a brief overview, drawing upon some of the salient features of the literature (both experimental and non-experimental) and with several examples from the Indian case, to make remarks relevant for policies in developing countries. First, a few words on definitional issues. We ll follow the standard economist s practice of defining corruption as use of public office for private gain (and thus leave out a lot of primarily inside-privatesector corruption, as in the many financial scams of recent years). We also need to keep a clear distinction between corrupt and illegal activities, since laws are widely different in different countries: for example, a lot of essentially corrupt ways in which money is legally used in influencing legislators and law-making in the US will be illegal in many other countries; as they say in India, in US corruption is in the process of making laws, in India it s mostly in breaking laws. In recent years in the Economics literature on measurement of corruption there has been a welcome move away from subjective measures (like those used by the widely-cited country rankings by Transparency International) to more objective measures, often directly from self-reported bribes in micro enterprise-level surveys. But the latter also have some problems: apart from any possible bias arising from using mainly the bribe-giver s point of view (except when direct observations like actual estimates by engineers, compared to reported expenditures, on Indonesian road projects are used as in Olken (2009)), most respondents may identify corrupt with only illegal activities. Also, they may ignore cases where a monetary bribe is not paid, like when connections are used to land a job or a contract, or where politicians may do a favor to you not in
exchange of money but some form of political support, or when officials steal your time rather than money (in the form of absenteeism or shirking). Corruption is usually presumed to originate in regulations and bureaucratic discretion, but even when that is the case, the solution does not always lie in the abolition of those regulations, particularly because we care about other outcomes (example: pollution, where a corrupt inspector is not enough reason for abolishing environmental regulations). Depending on the primary objective, one may even argue for some tolerance for corruption in specific cases. In the context of corruption originating in regulations, one puzzle that has been discussed in India is why corruption is perceived to have increased in the decades after economic liberalization and deregulation, when expectation was the opposite. One can suggest various explanations. (a) With economic growth public resources (like land, minerals, oil and gas fields, telecommunication spectrum) have shot up in market value, and hence their continued political allocation generates more corruption than before. (b) As elections become more expensive, illicit funding sources, in exchange of political favors, become more tempting. (c) With a larger role of the private sector and PPE s or publicprivate partnerships, regulatory agencies in different fields become more important; often weak and non-transparent regulations by these agencies increase opportunities for corruption. Also, as in many other countries, there is now more scope for officers privately employed post-retirement in sectors formerly regulated by them.
(d) Larger transfer payments, some of them relating to antipoverty programs, particularly when they are not universal, give scope for all kinds of fraudulent transactions, particularly in remote areas with weak institutions. (e) In the context of increasing decentralization and devolution of central funds, corruption may rise as collusion is easier to arrange at the local level. In areas of high social and economic inequality capture of local governments by the local elite is not uncommon. (f) Corruption may also be encouraged by the increasing ethnification of politics. This may happen for at least two reasons: (i) upwardly mobile, hitherto subordinate castes, lacking the well-developed networks nurtured by upper castes, may use money as a substitute for networks; (ii) dignity politics often trumps governance, so that many longoppressed caste groups keep on electing known corrupt members of their own group as symbolic gestures of group assertion. Some of the perception of increasing corruption may not, however, be real. With increasing education, political awareness, and active media more questions are raised about accountability and more of malfeasance may be exposed now, but it may not always be the case that corruption has actually increased compared to the past. Corruption may also genuinely increase when there is more public discussion of corruption, when everybody believes most people to be corrupt. Myrdal (1968) reported his discussion with Nehru about corruption, who told him: Merely shouting from the house-tops that everybody is corrupt creates an atmosphere of corruption.
People feel they live in a climate of corruption and they get corrupted themselves. In the theoretical literature there are a few models depicting this phenomenon of corruption as a frequency-dependent equilibrium. For a brief survey of this literature, see Bardhan(2005). When there are more people believed to be corrupt, at the margin the incentive for myself to be corrupt increases, with lower reputation loss when detected, lower chance of detection, lower search cost in finding a briber, etc. But ultimately the size of bribe is bid down by too many competing bribers. All this easily leads to multiple equilibria with different levels of corruption. Two otherwise similar countries (both in socio-economic structures and in moral attitudes) may end up with two very different equilibrium levels of corruption. Also initial conditions matter; a country may get locked into a corrupt equilibrium, which it may find difficult to escape. In situations of widespread belief of people being corrupt, intermediaries and touts have an interest in spreading cynical rumors and disinformation campaigns about officials (even when they are honest), which help the former in collecting bribes from their clients. In general the theoretical literature see, for example, Bose and Gangopadhyay (2009)-- has suggested that the presence of intermediaries, by lowering the uncertainty of whom and how much to bribe, lowering the chances of detection as well as of breach of corrupt deals, facilitate corruption and make anti-corruption policies more difficult to implement. This is also backed up by the experimental literature. In policy discussion economists often suggest tweaking the incentive structure, which includes both rewards and punishments.
In Singapore a wage premium above private-sector salaries has been reportedly successful (consistent with efficiency wage theory), but there is clearly a need for stringent monitoring and organizational changes to accompany such incentive reforms. Di Tella and Schargrodsky (2003) estimate in their analysis of micropanel data from public hospitals in Buenos Aires that a doubling of wages would cause a more than 20 per cent decline in the prices paid by hospitals in procuring basic supplies; but they also show that monitoring policies act as a complement to raising salaries in curbing procurement corruption. Similarly, Duflo et al(forthcoming) have found that financial incentives along with monitoring are likely to reduce teacher absenteeism in schools. The recent popular anti-corruption movement in India led by Anna Hazare, that attracted a great deal of attention, largely concentrated on demand for institutions of punishment of the bribe-taking official by a new bureaucracy (Jan Lok Pal) with the power to investigate and prosecute at all levels of officialdom. This would be certainly better than the existing Prevention of Corruption Act of 1988 in India, under which permission to prosecute has to be given by the Government (even though the Supreme Court has said it is not necessary). Experimental literature does suggest a strong deterrent effect of punishment. But a super-bureaucracy always raises questions like - Who monitors the monitor? - Do we have to bribe at yet another place now? China has the severest punishment (execution) with still rampant corruption. It is also often difficult to legally distinguish, beyond
reasonable doubt, between the outcome of simple ineptitude (or administrative failures) and that of dishonesty Then there is the question why the focus is on punishing the bribetaker, and not on the sometimes wealthier bribe-givers? If bribegivers continue to have the opportunity to make a lot of money, they ll find a way. It is thus more important to deal with the opportunities at the general level than to go for numerous punishments at the micro-level. Paradoxically, an increase in punishment may even increase the size of the bribe (to cover the rise in risk premium for the bribe-taker). It is also important to keep in mind that bribe-takers are not always individuals, it could be a political party (this is where the issue of reforming election finance becomes important). Even in offices, the bribe-collector may not be the ultimate bribe-taker, there may be an office-wide vertical network, sharing in the bribe. For policy purposes it is useful to distinguish between two different kinds of corruption: Speed money--the standard kind where you pay an official to speed up your file, you pay him to do what he is supposed to do anyway (Russians call this mzdoimstvo) Collusive corruption where you pay an official to do what he is not supposed to do (Russians call this likhoimstvo). Examples of the latter more insidious form of corruption: the official connives at or looks the other way when goods are smuggled, taxes are evaded, income or property value is under-assessed, driver s license or food ration card is issued to unqualified people, bids in public auctions are rigged, lower-quality materials are substituted in
government procurement, and so on. These cases involve collusion between the bribe-giver and the bribe-taker to evade laws, and both parties gain, thus neither is likely to report this to investigators. One solution to collusive corruption is giving the agent/bureaucrat high-powered incentives (like tax farming). But this may lead to extortion (as in the historical instances of tax farming). An independent authority to hear appeals against extortion (for example, over-assessment in taxes) is necessary, as has been introduced as part of tax reforms in some countries (like Mexico). Similarly, in fiscal devolution unconditional grants may be given to local governments, with random ex post audits. Mishra and Mookherjee (2012) have suggested an appropriate structure of fines (for example, lowering fines for the offense, say tax evasion) can handle the twin problems of collusion and extortion. In general some changes in administrative procedures may work better than erecting a new super-bureaucracy: One such change is reducing the monopoly power of the official that the client has to face (in getting, say, a death or marriage certificate or a passport), wherever feasible. In the US a citizen can get a passport from almost any post office, and so corruption on getting a passport is much less than in countries like India where the passport-issuing officer has monopoly. (I understand the state of Chhattisgarh in India has now outsourced issuing of some certificates to a whole range of authorized private agencies, centrally verified with computer technology).
Drugov (2010) shows that in cases where some bureaucrats are honest and some are not, competition may give increased incentive to the applicant to invest in the requisite qualification (learning to drive in the case of driver s license or carrying out pollution abatement in the case of a firm to be inspected). Of course, if the applicant remains unqualified, competition among bureaucrats may increase the chance of the applicant meeting an appropriately dishonest bureaucrat. Another case of reducing official monopoly power is from the historical examples of reducing corruption in the customs office in Singapore or the narcotics control section of the New York Police Department, by introducing overlapping jurisdictions of officers (in the latter case, for example, local, state, and federal agencies had overlapping involvement in controlling illegal drugs, which used to be a major source of police corruption). But these agencies have to be sufficiently independent, to minimize political collusion. This suggests that a demand in the Anna Hazare campaign for independence of the Central Bureau of Investigation (CBI) in India is very important. A less simple but important administrative reform for India is to change the current system of bureaucratic transfers and postings, which is a major source of illicit income of politicians in state secretariats. In general career promotion for officers in India depends more on seniority than on performance, so an officer has the incentive to maximize his or her loot in the short period of local posting before transfer. The Chinese governance system limits local official corruption by giving the local official more of a stake in the local economic performance. Chances of career promotion improve if the local area under his jurisdiction grows faster. So even when he
steals, he takes care, in his own self-interest, that the general economic performance of the area does not suffer too much. Mechanisms of public disclosure of information on demand (as envisaged in the landmark legislation on the right to information in India), availability of information on expenditure allocations and targets at the level of local governments, and use of technology (in tracking of public supplies, land records, etc.) are crucial. The empirical literature gives convincing evidence of the positive effects of availability of audit reports in Brazilian municipal elections--see Ferraz and Finan (2008) and of citizens report cards on politicians past performance and asset and criminal backgrounds in Delhi elections--- see Banerjee et al (2011). Experimental evidence also suggests that monitoring and ways of improving the perception of the probability of detection are effective. Some business people point out that in East Asia if you bribe someone, you know that the job will get done; in India, even after bribing you are never sure (possibly because of multiple veto powers in the bureaucracy). This system leads to special uncertainty and inefficiency of Indian corruption, in contrast with much of corruption in East Asia. Finally, one should be aware of the problems of pushing anticorruption policies too much. Overzealousness in erecting a corruption-detection machinery, for all the good intentions, can stifle honest and potentially beneficial but inherently risky decisions and dynamic leadership by public-sector officials. If a risky decision does not work out, and somebody somewhere is seen to gain as a result, that often works as prima facie evidence of corrupt intent of the official, and in anticipation an official will often opt for less bold
policies or status quo. This may lead to policy paralysis in an atmosphere of suspected pervasive corruption; it will be difficult to attract talent to administration under such circumstances. So while policies to change the system of incentives and punishments with the purpose of reducing corruption are welcome, one has to be careful that these policies do not seriously interfere with other important goals in the economy. References K. Abbink and D. Serra (2012), Anti-Corruption Policies: Lessons from the Lab, in D. Serra and L. Wantchekon (eds.), New Advances of Experimental Research on Corruption, Research in Experimental Economics, 15, 77-115. A. Banerjee and R. Pande (2009), Parochial Politics: Ethnic Preferences and Politician Corruption, Working Paper, Kennedy School, Harvard. G. Bose and S. Gangopadhyay (2009), Intermediation in Corruption Markets, Indian Growth and Development Review, 2(1), 39-55 R. Di Tella and R. Schargrodsky (2003), The Role of Wages and Auditing during a Crackdown on Corruption in the City of Buenos Aires, Journal of Law and Economics, 46(1), 269-92 M. Drugov (2010), Competition in Bureaucracy and Corruption, Journal of Development Economics, 92(2), 107-14
E. Duflo, R. Hanna, and S. Ryan (forthcoming), Monitoring Works: Getting Teachers to Come to School, American Economic Review C. Ferraz and F. Finan, Exposing Corrupt Politicians: The Effects of of Brazil s Publicly Released Audits on Electoral Outcomes, Quarterly Journal of Economics, 123(4), 703-45 A. Mishra and D. Mookherjee (2012), Controlling Collusion and Extortion: The Twin Faces of Corruption, unpublished, Boston University. B. Olken (2009), Corruption Perceptions vs. Corruption Reality, Journal of Public Economics, 93(7), 950-64